Fechar
Metadados

@MastersThesis{Montibeller:2018:AnEsCu,
               author = "Montibeller, Bruno",
                title = "An{\'a}lise espectro-temporal das culturas de milho, soja e 
                         cana-de-a{\c{c}}{\'u}car com dados de sensor OLI/Landsat-8",
               school = "Instituto Nacional de Pesquisas Espaciais (INPE)",
                 year = "2018",
              address = "S{\~a}o Jos{\'e} dos Campos",
                month = "2018-04-27",
             keywords = "espectro-temporal, an{\'a}lise interespec{\'{\i}}fica, 
                         an{\'a}lise intraespec{\'{\i}}fica, soja, milho, 
                         cana-de-a{\c{c}}{\'u}car, spectral-temporal, interspecific 
                         analysis, intraspecific analysis, soybean, corn, sugarcane.",
             abstract = "A agricultura, al{\'e}m de provedora de alimentos e energia, 
                         {\'e} uma das atividades econ{\^o}micas de maior destaque nas 
                         exporta{\c{c}}{\~o}es brasileiras. Produtos derivados da soja, 
                         do milho e da cana-de-a{\c{c}}{\'u}car s{\~a}o alguns exemplos 
                         de mercadorias agr{\'{\i}}colas exportadas. Assim, 
                         informa{\c{c}}{\~o}es relacionadas {\`a} {\'a}rea plantada e 
                         {\`a} produ{\c{c}}{\~a}o destas culturas s{\~a}o essenciais, 
                         uma vez que isso afeta a cota{\c{c}}{\~a}o destes produtos. A 
                         ado{\c{c}}{\~a}o de dados de sensoriamento remoto para pesquisas 
                         em {\'a}reas agr{\'{\i}}colas se apresenta como uma alternativa 
                         vi{\'a}vel e complementar aos m{\'e}todos atuais de 
                         levantamento, fornecendo informa{\c{c}}{\~o}es atualizadas e de 
                         forma cont{\'{\i}}nua. Essas pesquisas, geralmente, utilizam a 
                         informa{\c{c}}{\~a}o espectral como principal vari{\'a}vel para 
                         diversos objetivos como, mapeamento, estimativa de produtividade, 
                         c{\'a}lculo de biomassa, etc. Entretanto, as diferentes 
                         pr{\'a}ticas de manejo (com ou sem irriga{\c{c}}{\~a}o, 
                         {\'e}poca de plantio, etc.) empregadas no processo de 
                         produ{\c{c}}{\~a}o, podem alterar as caracter{\'{\i}}sticas 
                         espectrais das culturais. Neste sentido, o presente trabalho tem 
                         como hip{\'o}tese que a varia{\c{c}}{\~a}o espectral 
                         intraespec{\'{\i}}fica do milho, da soja e da canade- 
                         a{\c{c}}{\'u}car a partir de dados espectrais multitemporais do 
                         sensor OLI, {\'e} menor que a varia{\c{c}}{\~a}o 
                         interespec{\'{\i}}fica, independentemente das pr{\'a}ticas de 
                         manejo adotadas ao longo do seu desenvolvimento. Assim, o objetivo 
                         deste trabalho foi analisar o perfil espectro-temporal das 
                         culturas de soja, milho e cana-de-a{\c{c}}{\'u}car, cultivadas 
                         em uma {\'a}rea da Mesorregi{\~a}o de Campinas, SP, em dois anos 
                         agr{\'{\i}}colas completos (2014/2015 e 2015/2016), a partir de 
                         imagens do sensor OLI/Landsat-8, e verificar as diferen{\c{c}}as 
                         espectro-temporal das culturas. S{\'e}ries temporais de imagens 
                         OLI foram utilizadas para adquirir o perfil espectro-temporal das 
                         tr{\^e}s esp{\'e}cies cultivadas em {\'e}pocas distintas (safra 
                         e safrinha) e sob diferentes pr{\'a}ticas de manejos (irrigado e 
                         n{\~a}o irrigado). Foram utilizadas seis bandas espectrais 
                         (B2-B7) do sensor OLI e dois {\'{\i}}ndices de 
                         vegeta{\c{c}}{\~a}o (NDVI e EVI), os quais foram calculados a 
                         partir do valor de reflect{\^a}ncia das bandas. Com base na 
                         informa{\c{c}}{\~a}o espectro-temporal (bandas e 
                         {\'{\i}}ndices de vegeta{\c{c}}{\~a}o), uma an{\'a}lise 
                         hier{\'a}rquica de agrupamento foi realizada. A partir dos 
                         resultados, foi poss{\'{\i}}vel identificar que as bandas 
                         espectrais (B2-B7) se mostraram mais eficientes quando comparadas 
                         com os {\'{\i}}ndices de vegeta{\c{c}}{\~a}o, para a 
                         divis{\~a}o e gera{\c{c}}{\~a}o de clusters de cada cultura, 
                         permitindo assim, analisar a influ{\^e}ncia das pr{\'a}ticas 
                         culturais. Al{\'e}m disso, um conjunto de bandas espectrais, 
                         formado pelas tr{\^e}s bandas (B5-B6-B4), comumente utilizadas 
                         para interpreta{\c{c}}{\~a}o visual de alvos agr{\'{\i}}colas, 
                         geraram o mesmo resultado da an{\'a}lise baseada nas seis bandas. 
                         Com os resultados, {\'e} poss{\'{\i}}vel inferir que a 
                         varia{\c{c}}{\~a}o espectro-temporal intraespec{\'{\i}}fica 
                         {\'e} menor que a varia{\c{c}}{\~a}o espectro-temporal 
                         interespec{\'{\i}}fica, independentemente das pr{\'a}ticas de 
                         manejo empregadas. ABSTRACT: Agriculture, besides providing food 
                         and energy, is one of the most prominent economic activities in 
                         Brazilian exports. Products derived from soy, corn and sugar cane 
                         are some examples of exported agricultural commodities. Thus, 
                         information related to the acreage and the production of these 
                         crops are essential, since this affects the price of these 
                         products. The adoption of remote sensing data for research in 
                         agricultural areas is a viable and complementary alternative to 
                         current surveying methods, providing updated and continuous 
                         information. These surveys generally use spectral information as 
                         the main variable for several purposes such as mapping, forecast 
                         yield, biomass estimation, etc. However, different management 
                         practices (with or without irrigation, planting time, etc.) 
                         employed in the production process can alter the spectral 
                         characteristics of the crop. In this sense, the present work has 
                         as hypothesis that the intraspecific spectral variation of corn, 
                         soybean and sugarcane from multitemporal OLI sensor spectral data, 
                         is smaller than the interspecific variation, independently of the 
                         management practices adopted during crops development. Thus, the 
                         objective of this work was to analyze the spectrumtemporal profile 
                         of soybean, maize and sugarcane crops grown in an area of the 
                         mesoregion of Campinas, SP, in two full agricultural years 
                         (2014/2015 and 2015/2016 ), from OLI/Landsat-8 sensor images, and 
                         to verify the spectral-temporal differences of the cultures. Time 
                         series of OLI images were used to acquire the spectrum-temporal 
                         profile of the three species cultivated at different times 
                         (springsummer and summer-fall) and under different management 
                         (irrigated and nonirrigated). Six spectral bands (B2-B7) of the 
                         OLI sensor and two vegetation indices (NDVI and EVI) were used, 
                         which were calculated from the reflectance value of the bands. 
                         Based on the spectral-temporal information (bands and vegetation 
                         indices), a hierarchical cluster analysis was performed. From the 
                         results, it was possible to identify that the spectral bands 
                         (B2-B7) were more efficient when compared with the vegetation 
                         indexes, for the division and generation of clusters of each 
                         culture, thus, to analyze the influence of cultural practices. In 
                         addition, a set of spectral bands, formed by the three bands 
                         (B5-B6-B4) commonly used for visual interpretation of agricultural 
                         targets, generated the same result of the analysis based on the 
                         six bands. With the results, it is possible to infer that the 
                         intraspecific spectrum-temporal variation is smaller than the 
                         interspecific spectral-temporal variation, independently of the 
                         management practices employed.",
            committee = "Renn{\'o}, Camilo Daleles (presidente) and Sanches, Ieda Del'Arco 
                         (orientadora) and Luiz, Alfredo Jos{\'e} Barreto (orientador) and 
                         Ponzoni, Fl{\'a}vio Jorge and Aguiar, Daniel Alves de",
         englishtitle = "Spectrum-temporal analysis of the crops of corn, soy and sugar 
                         cane based on OLI/Landsat-8 sensor data",
             language = "pt",
                pages = "114",
                  ibi = "8JMKD3MGP3W34R/3QRM6DS",
                  url = "http://urlib.net/rep/8JMKD3MGP3W34R/3QRM6DS",
           targetfile = "publicacao.pdf",
        urlaccessdate = "23 nov. 2020"
}


Fechar